Analyze Content Style
POST /api/creative/hawken/style/analyze-content
Run the DE scene-scan against a scene and return the composed report.
Contract change (M5): the content form field now carries the
outline node ID (scene_id), not raw prose. The old LLM path passed
prose in; the deterministic path loads the scene’s prose internally
from its linked .docx document. expected_style is retained
for backwards compatibility but ignored — the scan reports whichever
register actually appears in the prose, not a target.
Request Body required
Section titled “Request Body required ”object
Scene id (not raw content)
Expected writing style (retained for API compat)
Responses
Section titled “ Responses ”Successful Response
Legacy name alias for :class:HawkenSceneScanResponse.
Kept because the M5 migration rewired /hawken/style/analyze-content
in place and the historical import surface should not break. New
callers should use :class:HawkenSceneScanResponse.
object
Deterministic style-consistency proxy (0-100)
Dominant register mapped to a style label
Classifier confidence (0-100)
Detected tone from sentiment + register
Suggested audience from register
Heuristic creativity score (0-100)
Heuristic engagement score (0-100)
Heuristic mechanics score (0-100)
Outline node id that was scanned
One finding produced by a sub-scanner during a scene scan (M5).
object
Source-prefixed rule identifier
Error | warning | info
Sub-report id that produced this finding
Human-readable diagnosis
Scope the finding is about
object
Sentiment (VADER) analysis sub-report for a scene (M5).
object
Positive-token density
Negative-token density
Neutral-token density
Overall compound [-1, +1]
Per-sentence compound scores in reading order
Tokens that matched the lexicon
Register (formal/casual/archaic) sub-report for a scene (M5).
object
Formal wordlist density
Casual wordlist density
Archaic wordlist density
Formal | casual | archaic | mixed | none
Quantitative scene-metric sub-report (M5).
object
Total words in scene
Total sentences
Words / sentence
Flesch-Kincaid grade level
Adverb-count / word-count
ProseGuard filter-word count / word-count
Per-sub-report status + count for the M5 scene scan.
Validation Error